toutvelu, a Regulator of Heparan Sulfate Proteoglycan Biosynthesis, Controls Guidance Cues for Germ-Cell Migration

Abstract

The primitive embryonic gonad in Drosophila melanogaster is composed of germ cells and somatic gonadal precursor cells (SGPs). The assembly of a functional gonad involves a complex series of germ-cell migration events, which are thought to be guided by attractive and repulsive cues. Here, we demonstrate a novel role for toutvelu (ttv), a regulator of heparan sulfate proteoglycan biosynthesis during this process. Germline clonal analysis suggests that maternal deposition of ttv is required for proper germ-cell migration. Conversely, ectopic expression of ttv in early embryos results in severe germ-cell migration defects and inappropriate spreading of Hh protein. Moreover, overexpression of ttv in only the receiving cells, rather than in the sending cells, leads to phenotypic consequences. Finally, supporting the claim that the signaling molecule Hedgehog (Hh) may function as a chemoattractant to guide germ cells, errant germ cells are found localized near pockets containing high concentrations of Hh protein.

IN Drosophila melanogaster, formation of primitive embryonic gonad involves recognition, attachment, and eventual coalescence of two cell types, namely germ cells and somatic gonadal precursor cells (SGPs) (Boyle and DiNardo 1995; Jaglarz and Howard 1995; Boyle et al. 1997; Moore et al. 1998). The germ cells, also termed pole cells, are formed at the posterior pole of the embryo under the control of the maternal determinant, oskar. The other component of the gonad, the SGPs, are formed in the dorsolateral mesoderm of parasegments 10–13. In contrast to the germ cells, SGP specification depends upon zygotically active patterning genes and inductive signals from the dorsal ectoderm (for a review, see Santos and Lehmann 2004a).

At the time of their formation, germ cells are situated outside of the embryo and need to be brought into physical proximity with the SGPs. Initially, through the movements of gastrulation, germ cells are carried inside the midgut pocket. By mid-embryogenesis, around stage 9–10, they begin to traverse the midgut epithelial wall, a process that appears to be governed, in part, by germ-cell autonomous factors such as tre-1 and stat (Kunwar et al. 2003; Li et al. 2003). The subsequent migration of the germ cells toward the SGPs is guided by the repulsive signals produced by the hindgut and chemoattractive signal(s) produced by the SGPs to inveigle the germ cells.

Earlier work has suggested that Drosophila HMGCoA reductase (Hmgcr) plays a critical role in this process (Van Doren et al. 1998). Hmgcr catalyzes the biosynthesis of mevalonate, a precursor for isoprenoids, and has been implicated in the production of an attractive signal by the SGPs (Santos and Lehmann 2004b). In hmgcr mutants, the germ cells linger after they traverse the midgut pocket and then scatter in the posterior of the embryo instead of migrating to the SGPs. Conversely, when hmgcr is ectopically expressed it induces germ-cell migration defects. While phenotypic analysis of hmgcr mutants indicates that it must play a critical role in a number of different biological pathways, we have recently shown (Deshpande and Schedl 2005) that one of its functions is to potentiate the release and/or transport of the Hedgehog (Hh) ligand from hh-expressing cells to the surrounding tissue (for a review of Hh signaling, see Ingham and McMahon 2001 and Lum and Beachy 2004). In hmgcr mutants, the Hh ligand is inappropriately retained in the sending cells. Conversely, when hmgcr is overexpressed in these same hh-expressing cells, it promotes the spreading of Hh protein through the adjacent tissue. The connection between hmgcr and hh signaling is of interest because previous studies have suggested that Hh, which is expressed in the SGPs, functions as one of the germ-cell chemoattractants. Like hmgcr, ectopic expression of hh leads to germ-cell migration defects. Consistent with the idea that germ cells respond to the Hh ligand, the two Hh receptors, Patched (ptc) and Smoothened (smo), were found to be required maternally for proper germ-cell migration. Moreover, as would be expected from their opposing roles in the hh pathway, germ cells compromised for ptc function clump prematurely whereas smo germ cells scatter randomly in the posterior of the embryo (Deshpande et al. 2001) While a role for Hh as a guidance molecule was rather unexpected, subsequent studies have supported the idea that morphogenetic signaling factors can moonlight as chemoattractants (Charron et al. 2003; Charron and Tessier-Lavigne 2005).

To gain further insight into the mechanisms that direct germ-cell migration and the role of the hh-signaling pathway in this process, we carried out a screen for germ-cell migration defects using deficiencies spanning the second chromosome. One of the loci that displayed germ-cell migration defects was toutvelu (ttv), which is known to be important for hh signaling. ttv encodes for an integral membrane protein belonging to the EXT gene family involved in the biosynthesis of heparan sulfate proteoglycans (HSPG). HSPGs are large macromolecules found in the extracellular matrix and are composed of a core protein that is sequentially modified with covalently linked glycosaminoglycan chains (for a detailed review, see Bernfield et al. 1999; Selleck 2001; Nybakken and Perrimon 2002; Lin 2004). Detailed biochemical and genetic characterization has revealed the sequential nature of the underlying enzymatic modifications during the HSPG biosynthesis. ttv, along with its two homologs, brother of totvelu (botv) and sister of toutvelu (sotv), together constitute the polymerases involved in this process. Furthermore, molecular genetic analysis of ttv function has suggested that it is needed for the distribution and/or stability of signaling molecules such as Hh, Wingless (Wg), tumor growth factor (TGF), and fibroblast growth factor (FGF) (Bellaiche et al. 1998; The et al. 1999; Bornemann et al. 2004; Han et al. 2004; Takei et al. 2004). Here we have investigated the involvement of ttv in germ-cell migration using both loss- and gain-of-function strategies.

MATERIALS AND METHODS

Immunohistochemistry:

The embryo stainings were performed essentially as described in Deshpande et al. (1995). Vasa and Hh antibodies are rabbit polyclonal antibodies. Both were used at a 1:500 dilution. Eyes Absent and Wingless antibodies are mouse monoclonal antibodies and were used at a 1:10 dilution. β-Galactosidase antibody was either a rabbit polyclonal purchased from Kappel (used at a 1:1000 dilution) or a mouse monoclonal antibody from the Developmental Hybridoma Bank (used at a 1:10 dilution).

Misexpression analysis:

The following UAS and GAL4 stocks were used for the misexpression studies: UAS-ttv, UAS- hmgcr, UAS-PtcΔloop2, hairy-GAL4, elav-GAL4, nanos-GAL4, patched-GAL4, UAS-βgalactosidase, and hh-GAL4/TM6 Ubx-lacz. In most experiments, males carrying two copies of the UAS-ttv transgene were mated with virgin females carrying two copies of the GAL4 transgene. The resulting progeny embryos were fixed and stained for subsequent analysis (Brand and Perrimon 1993).

RESULTS

ttv is required for germ-cell migration:

To identify new genes involved in germ-cell migration, we screened the stock center collection of second chromosomal deficiencies for deletions that, when homozygous, display defects in the migration of germ cells to the somatic gonad. One of the deficiencies that scored positive in the screen uncovered ttv. As with the deficiency, we found that embryos homozygous for a mutation in ttv also showed defects in germ-cell migration. While defects in migration were clearly evident in the deficiency or ttv mutant embryos, they were less severe than the defects reported for some of the other genes implicated in germ-cell migration, such as hmgcr (Van Doren et al. 1998). One reason why the migration defects in embryos lacking zygotic ttv activity are relatively modest is that there is a substantial maternal contribution of the ttv product. Consistent with this idea, embryos that are maternally and zygotically compromised for ttv activity display segmentation defects reminiscent of hh and wg pathway components, while embryos lacking only zygotic ttv activity show no segmentation defects, hatch, and survive until the pupal stage (The et al. 1999).

Since the severe segmentation defects in embryos lacking both maternal and zygotic ttv activity would make it difficult to interpret any abnormalities in germ-cell migration, we decided to test the involvement of ttv in germ-cell migration further in embryos that were maternally compromised but paternally rescued for ttv (ttv^M-Z+). Females carrying germline clones of ttv (Bellaiche et al. 1998) were mated with wild-type males or balancer males carrying a Ubx-lacz marker to minimize segmentation-related abnormalities. Embryos obtained from such females either were stained with Vasa antibody (not shown) or were double labeled with both β-galactosidase and Vasa antibodies (Figure 1). As illustrated by the examples in Figure 1, B and C, we found a range of germ-cell migration defects in stage 13 ttv^M-Z+ embryos. In some cases, a subset of the germ cells failed to coalesce (Figure 1B), while in other cases most of the germ cells were scattered (Figure 1C). While <5% of wild embryos have four or more “lost” germ cells, >40% of the ttv⁻ germline clone embryos have five or more “lost” germ cells (Table 1A).

Figure 1.—

Maternal deposition of ttv is required for proper germ-cell migration. Embryos obtained by mating females carrying ttv germline clones with balancer males carrying Ubx-lacZ were double labeled with Vasa (green) and β-galactosidase antibodies (red), respectively. Germ-cell migration defects induced by the maternal depletion ttv are not fully penetrant as seen by comparing embryos in A–C.

TABLE 1.

Germ-cell migration defects

	Germ cells lost
	0–4 (%)	5–6 (%)	7+ (%)
A. Embryos maternally compromised for ttv activity display varying degrees of germ-cell migration defectsa
ttv germline clone (N = 48)	58	15	27
Wild type (N = 49)	92	4	4
B. Expression of PtcΔloop2 in the germline disrupts germ-cell migrationb
nos-GAL4 (N = 122)	91	6	2
nos-GAL4; UAS-ptcΔloop2 (N = 100)	69	17	14

^aEmbryos obtained by mating females carrying ttv germline clones with wild type males were stained with anti-Vasa antibody and classified based on the number of germ cells that failed to coalesce with the somatic gonadal precursor cells.

^bFemales carrying nos-GAL4 were mated with UAS ptc Δloop2 males (X-linked transgene). In this scheme, female embryos received both the drive and the UAS transgene while male embryos received only the nos-GAL4 driver and consequently serve as an internal control. Embryos were co-immunostained with anti-Sxl and anti-Vasa antibodies. Germ-cell loss in Sxl-positive (female) embryos is considerably higher than that for Sxl-negative (male) embryos.

Table 1B also shows that the effects of depleting maternally derived ttv activity are quite similar to those observed when reception of the Hh signal in the migrating germ cells is disrupted by ectopic expression of a mutant form of the Patched protein, PtcΔloop2, using a germline-specific Nanos-GAL4 driver. PtcΔloop2 is a deleted form of Patched that cannot bind the Hh ligand but can still repress Smoothened (Smo). As a result, Hh-dependent signaling is constitutively repressed in the presence of this mutant. Ectopic expression of the PtcΔloop2 protein in migrating germ cells is therefore expected to interfere with the reception of the Hh signal and to partially mimic the smo⁻ phenotype (scattering of germ cells). As can be seen from Table 1, 31% of embryos carrying both transgenes (female) show loss of more than five germ cells whereas germ-cell migration in the control (male embryos) embryos is comparable to wild-type embryos. These findings would be consistent with the idea that smo activity is required in the germ cells for them to respond appropriately to the Hh ligand. Although we cannot completely exclude some type of cell autonomous role for ttv in the germ cells, we suspect, on the basis of ectopic expression experiments described below, that the defects in germ-cell migration seen in the ttv^M-Z+ embryos may be due to the absence of maternal ttv activity in the soma, rather than in the germ cells themselves. This view is also supported by the fact that germ-cell migration defects are evident in embryos lacking zygotic ttv activity.

Functioning of ttv in germ-cell migration:

Germ-cell migration is thought to be directed by attractive and repulsive cues that are produced by somatic gonadal precursor cells and hind gut, respectively (Zhang et al. 1997; Van Doren et al. 1998; Deshpande et al. 2001; Starz-Gaiano et al. 2001). One of the signaling molecules implicated in attracting the migrating germ cells to the somatic gonadal precursor cells is hh. Since ttv is known to be important in the transmission of the hh signal, it seemed likely that the effects of ttv mutations on germ-cell migration are due to compromised hh signaling. If this idea is correct, then it seemed possible that we might observe genetic interactions between ttv and a gain-of-function hh allele, hh^Mrt. Hh expression is not properly regulated in the hh^Mrt mutant and this misexpression induces minor but readily detectable germ-cell migration defects in heterozygous animals because the ectopic Hh from hh^Mrt competes with the Hh signal emanating from the somatic gonadal precursor cells (Deshpande and Schedl 2005). We have previously found that these minor germ-cell migration defects can be enhanced when hh^Mrt is trans to mutations in either dispatched (dsp) or hmgcr, which encode proteins involved in the release of the Hh ligand from the sending cells (Deshpande and Schedl 2005). Reducing the activity of these genes weakens the signal produced from the somatic gonadal precursor cells and makes it more difficult for the germ cells to distinguish between the correct signaling source and the competing ectopic hh^Mrt signaling sources. As would be expected if the effects of ttv on germ-cell migration are also due to a weakening of the Hh signal coming from the somatic gonadal precursor cells, ttv mutations dominantly enhance the germ-cell migration defects of hh^Mrt (data not shown).

To further investigate how ttv might modulate hh signaling to promote germ-cell migration, we next tested the effects of ectopically expressing Ttv protein. Because of its role in the synthesis of HSPGs for the extracellular matrix, we imagined that ttv would be required in somatic cells for the efficient transmission of the Hh signal rather than in the actual process of reception of the signal by the migrating germ cells. If this idea is correct, then it might be possible to perturb germ-cell migration by misexpression of ttv in somatic tissues. To explore this possibility, we used two different GAL4 drivers, hedgehog-GAL4 (hh-GAL4) and patched-GAL4 (ptc-GAL4). The hh-GAL4 driver was selected because it directs the expression of Ttv in cells that produce the Hh ligand while the ptc-GAL4 driver was chosen because it directs expression in cells that respond to the Hh ligand.

As can be seen in Figure 2, C and D, germ-cell migration in the UAS-ttv/hh-GAL4 embryos (stage 13–15) is not too much different from that seen in wild-type embryos or in embryos that carry only the UAS-ttv transgene (Figure 2, A and B; see also Table 2). In contrast, a quite different result is seen when expression of the UAS-ttv transgene is driven by ptc-GAL4. As evident in Figure 2, E and F, most of the germ cells are scattered in the posterior of the embryo and fail to coalesce with the SGPs. Severe perturbations in germ-cell migration were observed in UAS-ttv/ptc-GAL4 embryos from stage 13 onward and a significant number of germ cells failed to coalesce with the somatic gonadal precursor cells. Interestingly, in a subset of embryos, germ cells were clustered in the middle of the embryo. As a result, such embryos had three clusters of germ cells: two in the appropriate location on either side of the embryo in parasegment 10 and an additional cluster at a more anterior location (not shown; see below). To confirm this finding, we misexpressed Ttv protein using a wingless-GAL4 driver, which, like ptc-GAL4, drives expression in Hh-receiving cells. Table 2 shows that the effects on germ-cell migration with the wingless-GAL4 driver are similar to those seen for the ptc-GAL4 driver.

Figure 2.—

ttv overexpression in the Hh-receiving cells can induce germ-cell migration defects. Whole-mount staining of stage 13–15 embryos of the indicated genotype with antibodies against Vasa and β-galactosidase protein. Females carrying two copies of UAS-ttv were mated independently with the males of the genotype hh-GAL4/TM6 Ubx-lacz (A–D) or ptc-GAL4/UASβgal/ptc-GAL4 UASβgal (E and F). Embryos (6–12 hr old) were collected, fixed, and co-immunostained with β-galactosidase (red) and Vasa (green) antibodies. In the case of the hh-GAL4 driver, β-galactosidase-specific staining was used to identify the embryos of the correct genotype. The staining was visualized using confocal microscopy. As can be seen by comparing the panels, ttv overexpression is sufficient to induce germ-cell migration defects only when overexpressed using ptc-GAL4 (E and F). (A and B) UAS-ttv/Bal embryos. (C and D) Embryos of hh-GAL4/UAS-ttv genotype.

TABLE 2.

Ectopic expression of ttv in receiving cells demonstrates stronger germ-cell migration defects than ectopic expression in sending cells

No. of germ cells lost	Hedgehoga	Wingless
0–2	43 (68)	18 (35)
3–4	8 (13)	16 (31)
5–6	6 (9)	13 (25)
7+	6 (9)	4 (8)
N	63	51

UAS ttv females were mated with hh-GAL4 and wg-GAL4. Ectopic expression of ttv in sending cells, driven by hh expression, produces weaker germ-cell migration defects than ectopic expression of ttv in receiving cells, driven by wg expression. Numbers in parentheses are percentages.

^aData for the ectopic expression of ttv driven by hh are the sum of two independent experiments.

Ectopic expression of ttv in hh receiving alters the distribution of Hh:

The results reported above in Functioning of ttv in germ-cell migration indicate that misexpression of Ttv in Hh-producing cells has little effect on germ-cell migration, while misexpression in cells that should be receiving or responding to the Hh ligand had a dramatic effect on germ-cell migration. We wondered whether this difference was correlated with differences in the ability of Ttv to influence the stability, transmission, or localization of the Hh ligand when Ttv is ectopically expressed either in Hh-producing cells or in Hh-receiving cells. To answer this question, we examined the distribution of Hh protein in embryos in which ttv is expressed under the direction of the hh, wg, and ptc GAL4 drivers. As a control, we expressed Hmgcr using the same drivers. We have previously found that Hh distribution is altered when Hmgcr is ectopically expressed in cells that produce Hh, while there is no effect on distribution when Hmgcr is ectopically expressed in Hh-receiving cells.

As shown in Figure 3, when ttv expression is directed by the hh driver, the distribution of Hh protein in the stripes and interstripes resembles the wild type. Hh accumulates in a two-cell wide stripe in each parasegment while there is only a relatively low level of Hh protein on either side of this stripe. By contrast, when ttv expression is directed by the wg (not shown) or ptc (Figure 3) GAL4 drivers, the Hh gradient is greatly expanded and high levels of Hh protein extend from the two-cell stripe to near the middle of the interstripe region. Confirming our previous observations (Deshpande and Schedl 2005), when hmgcr expression is directed by the wg (not shown) or ptc (Figure 3) drivers, the distribution of Hh protein resembles that seen in wild-type embryos. A different result is obtained with the hh driver. Although the Hh parasegmental stripes are still discernible, the stripes are much broader than in wild type or when hmgcr expression is controlled by wg or ptc-GAL4 and there are high levels of Hh extending to almost the middle of the interstripe region (Figure 3).

Figure 3.—

ttv overexpression in Hh-receiving cells promotes the spread of Hh protein across the segment. Whole-mount staining of stage 11 embryos of the indicated genotype with antibodies against Hedgehog protein. Females carrying two copies of either UAS-hmgcr or UAS-ttv were mated independently with the males of the genotype hh-GAL4/TM6 Ubx-lacz or ptc- GAL4/ptc-GAL4. Embryos (6–12 hr old) were collected, fixed, and co-immunostained with β-galactosidase (mouse; not shown) and Hh (rabbit) antibodies. In the case of hh-GAL4 driver, a lack of β-galactosidase-specific staining was used to identify the embryos of the correct genotype. The staining was visualized using confocal microscopy and staining intensities were compared using identical settings. As can be seen by comparing the pattern of Hh antibody staining in each panel, Hmgcr is able to promote the transmission of the Hh ligand only when it is overexpressed using hh-GAL4 (sending cell) whereas ttv overexpression can promote Hh transmission only when overexpressed using ptc-GAL4 (receiving cell).

These findings indicate that ttv can promote the transmission of the Hh ligand when ectopically expressed in Hh-receiving cells, but not when expressed in cells that synthesize the Hh signal. These results are consistent with previous studies that have shown that in larval discs ttv is required for hh signaling in the receiving, but not the sending, cells (Bellaiche et al. 1998). Conversely, hmgcr can promote the transmission or movement of the Hh ligand only when it is ectopically expressed in cells that produce the Hh signal.

Overexpression of ttv alters germ-cell migration:

Next we wondered if overexpression of ttv in different regions of the embryo would have similar consequences on germ-cell migration. We decided to use drivers with varying tissue specificities such as elav-GAL4 (pan neural) and hairy-GAL4 (ectodermal). When expression of UAS-ttv was driven using a paternally derived X-linked elav-GAL4 transgene, we observed germ-cell migration defects in approximately half of the embryos (see Figure 4). To confirm that these embryos carried both the UAS-ttv transgene and the elav-GAL4 driver, we probed the embryos with both Vasa and antibody directly against the female-specific protein Sxl. As expected, female embryos (Sxl-positive embryos) displayed germ-cell migration defects whereas male (Sxl-negative embryos) did not (Figure 4; compare A and B to C and D). Interestingly, a significant number of the germ cells that did not coalesce with SGPs were found proximal to either the central nervous system (Figure 4A) or the ventral nerve cord (Figure 4B).

Figure 4.—

Ectopic expression of ttv using a nervous system driver leads to germ-cell migration defects. Embryos obtained by mating males carrying an X-linked insert of an elav-GAL4 transgene with females carrying UAS-ttv were co-immunostained using Sxl antibody (green) and anti-Vasa antibody (red). The Sxl antibody specifically labels female embryos, which carry both the transgenes. As expected, only the female embryos show significant germ-cell migration defects (A and B). The mispositioned germ cells are indicated with arrows. Also PS10, the two sites of gonad assembly are indicated with asterisks in A. In this experiment, male embryos, which lack Sxl protein, serve as a control (C and D).

Ectopic Ttv induces germ cells to congregate at sites of Hh protein accumulation:

To extend these results, we used another driver, hairy-GAL4, which is expressed in a segmental pattern in the epidermis. Embryos expressing Ttv in response to this driver exhibited a variety of germ-cell migration defects. In about half of the embryos, germ cells were seen in the vicinity of parasegments 10–12 but failed to coalesce with the SGPs (Figure 5A). Roughly a quarter of the total number of embryos showed germ cells that were scattered widely in the posterior of the embryo (Figure 5B). The remaining embryos displayed a “pseudogonad”-like assembly in the middle of the embryo (Figure 5C).

Figure 5.—

Ectopic expression of ttv using an epidermal driver leads to germ-cell migration defects. hairy-GAL4 females were mated with males carrying a UAS-ttv transgene. The embryos collected from these females were stained with Vasa antibody, which specifically labels germ cells, and the staining was detected using diaminobenzidine. (A–C) As discussed in the text, different germ-cell migration phenotypes are evident in embryos ectopically expressing Ttv.

We considered two different explanations for these defects. In the first, hairy-GAL4-dependent overexpression of Ttv induces the formation of SGPs at ectopic locations in the embryo and these cells then attract a subset of the germ cells, in some cases giving rise to pseudo-coalesced gonads in the middle of the embryo like that in Figure 5C. In the second possible explanation, misexpression of Ttv is able to promote the accumulation of high levels of Hh protein at ectopic sites (by influencing its transport and/or stability) and these pockets of concentrated Hh protein in turn attract the germ cells. The presence of multiple sources of concentrated Hh would tend to “confuse” the germ cells and would explain why they fail to coalesce with the SGPs. Since the germ cells might also migrate toward these pockets of Hh protein, this would also explain the scattering phenotype (Figure 5B) and the formation of “pseudogonads” at ectopic locations (Figure 5C).

If the first model is correct, the pseudo-coalesced gonads should contain Eyes-Absent-positive SGP cells. However, we did not observe any Eyes-Absent-positive SGPs at ectopic locations in these embryos. Furthermore, we did not find any evidence of disruption in the assembly of the somatic gonad at its normal location (not shown). To test if pockets of concentrated Hh protein might be responsible for disrupting germ-cell migration, we double labeled the embryos using Vasa and Hh antibodies. Figure 6, A–I, shows images captured at three different planes of a single stage 14 embryo that exhibits the “scattered” germ-cell migration phenotype (as in Figure 5B). As expected from the second model, we find that the pattern of Hh accumulation differs from wild type and that there are many pockets of concentrated Hh protein throughout the embryo (see arrows in Figure 6). Supporting the idea that the germ cells might be attracted to these pockets of concentrated Hh, the mislocalized germ cells often (although not always) cluster near sites that have a high concentration of Hh protein (arrows in Figure 6 indicate areas of high Hh concentration and nearby germ cells).

Figure 6.—

Mislocalization of Hh protein is correlated with defects in germ-cell mispositioning. Hairy-GAL4/UAS-ttv embryos were double labeled with Vasa (green) and Hedgehog antibodies (red), respectively. (A, D, and G) Merged images showing various mislocalized germ cells clustering proximal to ectopic sources of Hh protein. The areas where Hh ligand is accumulated are indicated by arrows. Shown are three different planes from a single embryo. (B, E, and H) Mislocalization of Hh protein induced by ectopic expression of Ttv in the same planes as A, D, and G. (C, F, and I) Mislocalized germ cells in the same plane as A, D, and G.

DISCUSSION

In the studies reported here we have investigated the role of Ttv in the process of germ-cell migration. We initially identified ttv as a migration “factor” in a genetic screen for zygotic mutations that perturb germ-cell migration. Since there is a substantial maternal contribution of Ttv that is able to sustain development until the pupal stage, this would suggest that the process of migration may be more sensitive to the levels of Ttv activity than other aspects of development. Consistent with this idea, germ-cell migration defects are also evident when maternal Ttv activity is depleted using germline clones, even though again there are no obvious abnormalities in the patterning of these embryos. While we cannot exclude a role for ttv in the germ cells themselves, it seems likely that ttv activity is required in the soma, not the germline. First, germ cells are transcriptionally quiescent until just prior to their association with the SGPs and must depend upon maternally contributed gene products much longer than the zygotically active somatic cells. For this reason, one would not expect germ cells to be particularly sensitive to a zygotic loss of ttv activity, especially given that the maternal contribution of ttv seems to be quite substantial and is capable of sustaining apparently normal somatic development until the pupal stage (The et al. 1999). Second, Ttv functions in the biosynthesis of the extracellular matrix. Since this matrix provides the substratum for the transmission of signaling molecules and for cell movement, it seems likely that the somatic tissues through which the germ cells migrate are likely to be responsible for the synthesis of this matrix, not the germ cells. Also consistent with the idea that Ttv activity is required in the soma, ectopic expression of Ttv in somatic tissues, including the epidermis and CNS, induces germ cells to migrate toward sites distant from the coalescing SGPs.

The finding that ectopic Ttv promotes the transmission/accumulation of Hh protein in embryos is in line with previous studies that demonstrated a role for Ttv in the receiving cells in the hh signaling pathway (Bellaiche et al. 1998; Gallet et al. 2003). Recently, however, ttv has also been implicated in modulating the activity of other signaling molecules including Wg and FGF (reviewed in Lin 2004). Indeed, as would be expected if ttv functions in other signaling pathways in addition to hh, we found that the distribution of Wg protein appeared to be altered in embryos in which ttv was ectopically expressed using ptc, wg, and hh drivers. These observations raise the question of whether the alterations in germ-cell migration induced by Ttv misexpression or reduced Ttv activity are due to effects on the transmission/accumulation of Hh or some other signaling molecule(s). While we have not been able to detect any influence of wg on germ-cell migration, this does not exclude the possibility that the effects of ttv on germ-cell migration may be due to some other unidentified signaling molecules. On the other hand, several lines of evidence are consistent with the idea that ttv influences germ-cell migration because of its role in the transmission of the Hh signal in receiving cells. First, like other components of the hh-signaling pathway, ttv dominantly enhances the very modest germ-cell migration defects of the hh^Mrt allele. Second, germ-cell migration defects are induced when ttv is overexpressed in hh-receiving cells, but not in hh-sending cells. This specificity for the receiving cell would be expected from the known role of ttv in hh signaling. Third, in embryos misexpressing ttv germ cells are induced to migrate toward and often accumulate near sites that have abnormally high concentrations of Hh protein. While these observations argue that ttv likely promotes germ-cell migration toward the SGPs by potentiating the Hh signal emanating from these cells, it remains to be determined how Ttv protein actually enhances the transmission of the Hh ligand through the mesoderm.